JPWO2007060780A1 - Idea support device, idea support method, and idea support program - Google Patents

Abstract

The idea support device stores a text database that stores a plurality of texts, a text mining unit that analyzes the plurality of texts stored in the text database and outputs a text sentence as a result, and a conversion keyword. A keyword set database that extracts keywords from the text sentence using conversion keywords stored in the keyword set database; and the text sentence is extracted by the keyword extraction section in the sentence. A keyword conversion unit that converts the keyword into a conversion keyword stored in the keyword set database, and a result output unit that outputs a text sentence converted by the keyword conversion unit.

Description

  The present invention relates to an idea support device, an idea support method, and an idea support program that support the idea of a new idea.

  There is 4W1H conversion as a method to support the idea of a new idea.

  In this method, the value provided by the existing service is prepared in the form of text, keywords corresponding to Who, Why, Where, When and How are extracted from the text, and the extracted keyword is a keyword different from that keyword. To create new text. This created text supports ideas of unprecedented value and is used to identify potential needs.

  On the other hand, the text mining system performs dependency analysis on the text in the database, recognizes the structure of the sentence, extracts frequent patterns based on the appearance frequency of the partial structure of the sentence, The pattern to be output as the mining result. Therefore, the text mining system can extract sentences and keywords that characterize the database.

FIG. 1 is a block diagram illustrating an example of a text mining system. In FIG. 1, the text mining system includes a text DB 101, a text analysis unit 102, a similar structure generation unit 103, a frequent pattern detection unit 104, a result output unit 105, and a keyword set DB 106. An example of a conventional text mining system is described in Patent Document 1.
JP 2004-246491 A

  The first problem is that in the case of idea support using 4W1H conversion, humans must prepare text before keyword conversion, extract keywords to be converted, and convert the extracted keywords. . The reason is that the method of preparing the text, the method of extracting keywords, and the method of conversion are generally not obvious.

  The second problem is that it is difficult to use the text mining system for idea support. This is because the method of using text mining for idea support is generally not self-evident.

  An object of the present invention is to provide efficient idea support.

  In order to achieve the above object, an idea support apparatus according to the present invention analyzes a text database storing a plurality of texts and a plurality of texts stored in the text database by text mining, and outputs the resulting text. A text mining unit, a keyword set database for storing conversion keywords, a keyword extraction unit for extracting keywords from the text of the mining results using the conversion keywords stored in the keyword set database, and the mining result For text, a keyword conversion unit that converts a keyword extracted by the keyword extraction unit in a sentence to a conversion keyword stored in the keyword set database, and outputs the text converted by the keyword conversion unit result And a radical 19.

  The idea support method according to the present invention is an idea support method performed by an idea support apparatus including a text database for storing a plurality of texts and a keyword set database for storing conversion keywords, which is stored in the text database. A text mining step of analyzing the plurality of texts obtained by text mining and outputting the resulting text; a keyword extraction step of extracting keywords from the text using conversion keywords stored in the keyword set database; A keyword conversion step of converting the extracted keyword in the sentence into a conversion keyword stored in the keyword set database, and a result output step of outputting the converted text with respect to the text of the mining result Equipped with a.

  According to the above invention, the text mining result characterizing the text database is used as the text into which the keyword is converted, and the text having a different meaning from the text mining result is automatically generated by the keyword conversion. This generated text will help you come up with new ideas.

  For this reason, it is possible to automate the idea support, and it is possible to provide efficient idea support. In addition, text mining can be used for idea support.

  The idea support device includes a collation unit that collates the text converted by the keyword conversion unit and the text in the text database, and ranks the converted text based on the collation result. In addition, it is preferable that the result output unit outputs the converted text ranked by the collation unit.

  According to the above-described invention, it is possible to rank the converted text based on the difference in contents between the text into which the keyword is converted and the text in the text database.

  For this reason, for example, it is possible to give a high ranking to the converted text that is semantically different from the text in the text database and is likely to support a new idea. Therefore, for example, when the text into which the keyword is converted is rearranged in the ranking, the user can easily find the text that can support the new idea.

  The keyword set database preferably stores synonyms, synonyms or antonyms used in the text mining unit as the conversion keywords.

  According to the above invention, synonyms, synonyms or antonyms used in text mining can also be used as conversion keywords.

  Preferably, the keyword set database stores a plurality of conversion keyword candidates, and the keyword conversion unit uses the conversion keyword candidates associated with the text database in advance as the conversion keywords.

  According to the above invention, the conversion keyword is switched according to the text stored in the text database. For this reason, it becomes possible to convert the text of the mining result using the conversion keyword that is most suitable for the text stored in the text database.

  In addition, it is preferable that the collation unit assigns a higher rank to the text converted by the keyword conversion unit as the frequency of appearance in the text database decreases.

  According to the above-described invention, it is possible to give a high ranking to the converted text that is likely to support a new idea that is semantically different from the text in the text database.

  The idea support program of the present invention is an idea support program for causing a computer connected to a text database storing a plurality of texts and a keyword set database storing conversion keywords to execute an idea support process. A plurality of texts stored in the text database is analyzed by text mining, and the result is output using the text mining process and the conversion keywords stored in the keyword set database. A keyword extraction process for extracting a keyword from the text, and a keyword conversion process for converting the extracted keyword in a sentence into a conversion keyword stored in the keyword set database for the mining result text; A result output process for outputting the serial converted text, the ideas assistance process comprising, causes the computer to perform.

  According to the invention described above, the idea support method can be executed by the computer.

  According to the present invention, idea support using text mining can be automated. The reason is that a mining result characterizing the text database is used as the text to which the keyword is converted, and text having a meaning different from the mining result is automatically generated by the keyword conversion.

FIG. 1 is a block diagram showing a configuration of a conventional example. FIG. 2 is a block diagram showing the configuration of the best mode for carrying out the first invention of the present invention. FIG. 3 is a flowchart for explaining the operation of the best mode for carrying out the first invention. FIG. 4 is a block diagram showing a specific example of the best mode for carrying out the first invention. FIG. 5 is an explanatory diagram showing a specific example of the conversion keyword. FIG. 6 is an explanatory diagram showing a specific example of the conversion keyword.

Explanation of symbols

DESCRIPTION OF SYMBOLS 201 Text database 202 Text mining part 203 Keyword extraction part 204 Keyword conversion part 205 Collation part 206 Result output part 207 Keyword set database

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  FIG. 2 is a block diagram showing an idea support device according to an embodiment of the present invention.

  In FIG. 2, the idea support device includes a text database 201, a text mining unit (hereinafter referred to as “mining unit”) 202, a keyword extraction unit 203, a keyword conversion unit 204, a keyword set database 207, and a collation. Unit 205 and result output unit 206.

  The text database 201 stores a plurality of texts. The mining unit 202 analyzes a plurality of texts stored in the text database 201 by text mining. “Text mining” is also generally referred to as “data mining”. Hereinafter, these are also referred to as “mining”.

  The keyword set database 207 stores conversion keywords. For example, the keyword set database 207 stores keywords such as synonyms, synonyms and antonyms used when the mining unit 202 performs mining as conversion keywords.

  The keyword extraction unit 203 extracts keywords from the text that is the mining result. Specifically, the keyword extraction unit 203 uses the conversion keywords stored in the keyword set database 207 to extract keywords from the text that is the mining result.

  The keyword conversion unit 204 converts the keyword extracted by the keyword extraction unit 203 in the text as the mining result into a conversion keyword stored in the keyword set database 207.

  The collation unit 205 collates the text converted by the keyword conversion unit 204 with the text in the text database 201, and ranks the converted text based on the collation result. For example, the collation unit 205 assigns a high rank (small number) to text that does not exist in the text database 201.

  The result output unit 206 outputs the text converted by the keyword conversion unit 204. In other words, the result output unit 206 outputs the text to which the ranking is given by the collation unit 205.

  Note that the mining unit 202, the keyword extraction unit 203, the keyword conversion unit 204, and the collation unit 205 may be realized by a computer that executes a program recorded in a recording medium such as a computer-readable memory.

  In addition, the idea support device includes a text database 201, a mining unit 202, a keyword extracting unit 203, a keyword converting unit 204, a keyword set database 207, a matching unit 205, and a result output unit 206. It may be realized as a support system.

  Next, an example of the operation of this idea support device will be described with reference to FIGS. 2 and 3. FIG. 3 is a flowchart for explaining an example of the operation of the idea support apparatus.

  In the following, in the keyword set database 207, a Who conversion keyword set in which keywords corresponding to Who are collected, a Why conversion keyword set in which keywords corresponding to Why are collected, and a keyword corresponding to Where are collected. It is assumed that a Where conversion keyword set, a When conversion keyword set in which keywords corresponding to When are collected, and a How conversion keyword set in which keywords corresponding to How are collected are stored.

  Further, the keyword extraction unit 203 refers to the keyword set database 207 and extracts keywords corresponding to 4W1H (Who, Why, Where, When, How) from the text that is the text mining result.

  When a plurality of texts are input to the text database 201 in step 301, the mining unit 202 analyzes the text stored in the text database 201 by text mining in step 302 and outputs the text that is the text mining result. To do.

  The mining unit 202 provides the keyword extraction unit 203 with text that is the result of text mining. This text is given a tag such as part of speech.

  The keyword extraction unit 203 executes Step 303 when receiving the text of the text mining result.

  In step 303, the keyword extraction unit 203 extracts a keyword corresponding to 4W1H (Who, Why, Where, When, How) from the text that is the text mining result, and the extracted keyword and the text mining result. The text is provided to the keyword conversion unit 204.

  When the keyword conversion unit 204 receives the extracted keyword and the text that is the text mining result, the keyword conversion unit 204 executes Step 304.

  In step 304, the keyword conversion unit 204 converts the extracted keyword with reference to the keyword set database 207.

  Specifically, the keyword conversion unit 204 converts the keyword extracted by the keyword extraction unit 203 in the text that is the text mining result into a conversion keyword stored in the keyword set database 207.

  Furthermore, the keyword conversion unit 204 converts the keyword corresponding to Who extracted by the keyword extraction unit 203 into another keyword in the Who conversion keyword set, and is extracted by the keyword extraction unit 203. The keyword corresponding to Why is converted into another keyword in the keyword list for Why conversion.

  Further, the keyword conversion unit 204 converts the keyword corresponding to “Where” extracted by the keyword extraction unit 203 into another keyword in the “Where” conversion keyword set, and the When extracted by the keyword extraction unit 203. The keyword corresponding to is converted to another keyword in the When conversion keyword set, and the keyword corresponding to How extracted by the keyword extraction unit 203 is converted to another keyword in the How conversion keyword set. To do.

  The keyword conversion unit 204 combines these conversions to generate a plurality of texts with converted keywords. The keyword conversion unit 204 provides the collation unit 205 with the text converted from the keyword.

  When the collation unit 205 receives the text with the converted keyword, it executes Step 305.

  In step 305, the collation unit 205 collates the text converted by the keyword conversion unit 204 with the text in the text database 201, and assigns a higher rank (small number) to the converted text that does not exist in the text database 201. To do. The collation unit 205 provides the converted text to which the rank is assigned to the result output unit 206.

  When the result output unit 206 receives the converted text to which the rank is assigned, the result output unit 206 executes Step 306.

  In step 306, the result output unit 206 outputs the converted text with the ranking. For example, the result output unit 206 displays the post-conversion text that has been ranked.

  In this example, the collation unit 205 assigns a high rank to the converted text that does not exist in the text database 201, but, for example, the one that appears in the text database 201 with a low frequency using a statistical technique. You may order in order.

  In addition, the matching unit 205 uses a thesaurus to define the distance between the conversion source keyword and the converted keyword, and weights the converted keyword according to the distance between the conversion source keyword and the conversion source keyword. And ranking may be given to the converted text based on the weighting.

  According to this embodiment, the following effects can be obtained.

  The first effect is that idea support is automated. This is because the text of the mining result obtained from the text database 201 is used as an existing idea, and a keyword having a different idea can be automatically generated by converting a keyword in the existing idea into another keyword. .

  The second effect is that the efficiency of idea support is improved. The reason is that a mining result characterizing the text database 201 is used as a keyword conversion target, and the collation unit 205 collates the keyword conversion result with the text database 201 and ranks the conversion results based on the collation result. This is because it becomes possible to preferentially present something that is likely to lead to a new idea.

(Example)
Next, this embodiment will be described using specific examples.

  FIG. 4 is a block diagram showing an example of the embodiment of the present invention. In FIG. 4, the same components as those shown in FIG.

  FIG. 5 is an explanatory diagram showing an example of conversion keywords stored in the keyword set database 207.

  In FIG. 5, the conversion keywords “mature”, “young” and “elderly” are associated with Who, and the conversion keywords “male” and “female” are associated with Who. ing.

  Hereinafter, each of “mature group”, “young group”, and “elderly person” associated with Who and “male” and “female” associated with Who are referred to as a conversion keyword set.

  Here, it is assumed that reputation information of passenger cars is handled.

  The text database 201 stores reputation information of a plurality of passenger cars as text. In this case, the domain of the text stored in the text database 201 is “passenger car reputation information”.

  It is assumed that the mining unit 202 performs text mining on the reputation information of a plurality of passenger cars stored in the text database 201 and obtains a result 401 that “A company's luxury car is for mature men”. The result 401 becomes a target sentence for 4W1H conversion.

  The keyword extraction unit 203 extracts “middle class” and “male” 402 corresponding to Who from the mining result 401 as keywords. For example, the keyword extraction unit 203 matches the keywords (“mature”, “young” or “elder” and “male” or “female”) associated with Who in the keyword set database 207. Are extracted from the mining result 401 as keywords corresponding to Who.

  The keyword conversion unit 204 refers to the keyword set database 207, converts the keywords extracted by the keyword extraction unit 203 in the mining result 401 into conversion keywords, and generates a plurality of conversion results 403.

  Specifically, the keyword conversion unit 204 refers to the keyword set database 207 to change the “mature group” in the mining result 401 to “young group” and “elderly group” associated with the “middle class”. The converted text (conversion result 403) “the luxury car of company A is for young men” and “the luxury car of company A is for elderly men” are generated.

  Also, the keyword conversion unit 204 refers to the keyword set database 207 to convert “male” in the mining result 401 to “female” associated with “male”, and converts the converted text “Company A” "Luxury cars are for mature women".

  Further, the keyword conversion unit 204 refers to the keyword set database 207 to convert the “mature group” in the mining result 401 into the “young group” associated with the “mature group”, and further, the mining result 401 The “male” inside is converted into “female” associated with “male”, and the converted text “A company's luxury car is for young women” is generated.

  Further, the keyword conversion unit 204 refers to the keyword set database 207 to convert “mature group” in the mining result 401 to “elderly” associated with the “middle class”, and further, the mining result 401 The “male” inside is converted into “female” associated with “male”, and the converted text “A company's luxury car is for elderly women” is generated.

  Here, the conversion results 403 are generated by the number of keyword combinations.

  For this reason, the collation unit 205 collates each conversion result 403 with the text in the text database 201, for example, ranks the conversion results 403 in order from the least frequent conversion results 403 that appear in the text database 201, and arranges the conversion results 403. Change. As a result, for example, a collation result 404 “A company's luxury car is for young women” is obtained with high priority.

  Since priority is given to the collation result 404 that does not exist in the text database 201, there is a possibility that it becomes a new idea (latent needs).

  As the conversion keyword set stored in the keyword set database 207, for example, a set of keywords such as synonyms, synonyms and antonyms used in the mining unit 202 may be used. In this case, synonyms, synonyms or antonyms used in mining can also be used as conversion keywords.

  The conversion keyword set can be dynamically switched according to the domain to be mined.

  For example, the keyword set database 207 originally has keywords (conversion keyword candidates) such as “male”, “female”, “young”, and “elderly”, and these keywords are associated with domains in advance. It is assumed that One keyword may be associated with a plurality of domains. In this case, when a domain is determined, a keyword associated with the domain is determined. This correspondence relationship is registered in advance in the keyword conversion unit 204, for example.

  Therefore, if a set composed of keywords associated with a domain is used as a set of conversion keyword candidates for that domain, the optimal conversion keyword set for the domain can be dynamically switched. It becomes possible.

  For example, the keyword conversion unit 204 accepts a domain of the text database 201 and uses a keyword associated with the domain in advance as a conversion keyword.

  In this case, the conversion keyword is switched according to the text stored in the text database 201. For this reason, it becomes possible to convert the text using the conversion keyword most suitable for the text stored in the text database 201.

  The keyword set shown in FIG. 5 corresponds to “passenger car reputation information” (domain), and the keyword set shown in FIG. 6 corresponds to “mobile phone reputation information” (domain). Comparing the two, the keyword “elderly” belongs to different sets.

  Therefore, when dealing with "reputation information on passenger cars", the keyword "elderly" is converted to "mature" and "young", but when dealing with "reputation information on mobile phones" The same keyword “elderly” is converted to “high school girl”, “OL”, and “business man”. For this reason, more efficient idea support is possible.

  In the present embodiment, only the Who among the 4W1H conversion target keywords has been described. However, different keywords (Why, Where, When, How) are similarly converted. Also, the conversion target keyword is not limited to 4W1H, but may be 5W1H to which What is added, for example. When the conversion target keyword is 5W1H, the same processing as described above is possible.

  According to the present embodiment, the mining result characterizing the text database 101 is used as the text into which the keyword is converted, and the keyword conversion performed by the keyword extracting unit 203 and the keyword converting unit 204 has a different meaning from the mining result. The text you have is automatically generated. This generated text helps to come up with new ideas.

  For this reason, it is possible to automate the idea support, and it is possible to provide efficient idea support. In addition, text mining can be used for idea support.

  In the present embodiment, the collation unit 205 collates the text converted by the keyword conversion unit 204 with the text in the text database 201, and ranks the converted text based on the collation result. .

  In this case, the converted text can be ranked based on the difference in content between the text in which the keyword is converted (the converted text) and the text in the text database 201. For this reason, for example, it is possible to assign a high ranking to converted text that is semantically different from the text in the text database 201 and can support a new idea.

  Therefore, when the text in which the keyword is converted (the converted text) is rearranged according to the ranking, the user can easily find the converted text that can support the new idea.

  Further, in this embodiment, the collation unit 205 assigns a higher rank to the text converted by the keyword conversion unit 204 as the frequency of appearance in the text database 201 decreases.

  In this case, it is possible to give a high ranking to the converted text that is semantically different from the text in the text database 201 and can support a new idea.

  In the embodiment described above, the illustrated configuration is merely an example, and the present invention is not limited to the configuration.

The present invention can be applied to applications such as idea support at the time of product planning or strategy planning and the discovery of latent needs.

Claims (15)

A text database that stores multiple texts;
A text mining unit that analyzes a plurality of texts stored in the text database by text mining, and outputs the resulting text;
A keyword set database for storing conversion keywords;
A keyword extraction unit that extracts keywords from the text of the mining result using conversion keywords stored in the keyword set database;
For the text of the mining result, a keyword conversion unit that converts a keyword extracted by the keyword extraction unit in a sentence into a conversion keyword stored in the keyword set database;
An idea support apparatus comprising: a result output unit that outputs the text converted by the keyword conversion unit. The text further converted by the keyword conversion unit and the text in the text database are collated, and further includes a collation unit that ranks the converted text based on the collation result,
The idea support apparatus according to claim 1, wherein the result output unit outputs the converted text ranked by the collation unit.   The idea support apparatus according to claim 1, wherein the keyword set database stores synonyms, synonyms or antonyms used in the text mining unit as the conversion keywords. The keyword set database stores a plurality of conversion keyword candidates,
The idea support device according to any one of claims 1 to 3, wherein the keyword conversion unit uses, as the conversion keyword, the conversion keyword candidate associated in advance with the text stored in the text database. .   The idea according to any one of claims 2 to 4, wherein the collation unit assigns a higher rank to the text converted by the keyword conversion unit as the frequency of appearance in the text database decreases. Support device. An idea support method performed by an idea support device including a text database storing a plurality of texts and a keyword set database storing conversion keywords,
A text mining step of analyzing a plurality of texts stored in the text database by text mining and outputting the resulting text;
A keyword extraction step of extracting a keyword from the text of the mining result using a conversion keyword stored in the keyword set database;
A keyword conversion step of converting the extracted keyword in the sentence into a conversion keyword stored in the keyword set database for the mining result text;
A concept output method comprising: a result output step of outputting the converted text. A matching step of matching the converted text with text in the text database and ranking the converted text based on the matching result;
The idea support method according to claim 6, wherein in the result output step, the converted text with the ranking is output.   The idea support method according to claim 6 or 7, wherein the keyword set database stores synonyms, synonyms or antonyms used in the text mining step as the conversion keywords. The keyword set database stores a plurality of conversion keyword candidates,
The idea support method according to any one of claims 6 to 8, wherein, in the keyword conversion step, the conversion keyword candidates associated in advance with the text stored in the text database are used as the conversion keywords. .   The idea support method according to any one of claims 7 to 9, wherein, in the collation step, the converted text is given a higher rank as the frequency of appearance in the text database is lower. An idea support program for causing a computer connected to a text database storing a plurality of texts and a keyword set database storing conversion keywords to execute an idea support process,
A text mining process for analyzing a plurality of text stored in the text database by text mining and outputting the resulting text;
A keyword extraction process for extracting a keyword from the text of the mining result using a conversion keyword stored in the keyword set database;
For the mining result text, a keyword conversion process for converting the extracted keyword in a sentence into a conversion keyword stored in the keyword set database;
An idea support program for causing the computer to execute an idea support process including a result output process for outputting the converted text. The idea support process further includes a collation process that collates the converted text with the text in the text database, and ranks the converted text based on the collation result,
12. The idea support program according to claim 11, wherein in the result output process, the converted text with the ranking is output.   The idea support program according to claim 11 or 12, wherein the keyword set database stores synonyms, synonyms or antonyms used in the text mining process as the conversion keywords. The keyword set database stores a plurality of conversion keyword candidates,
14. The idea support device according to claim 11, wherein, in the keyword conversion process, the conversion keyword candidates associated in advance with the text stored in the text database are used as the conversion keywords. program.   The idea support program according to any one of claims 12 to 14, wherein, in the collation process, the converted text is given a higher rank as the frequency of appearance in the text database decreases.

Images